It has proved difficult in the past to obtain properly representative spectra of paste or paste-like materials, because of problems in forming a uniformly thin film of the paste. In one known technique a thin film is pressed out between a pair of crystal plates and analysed in transmission. This technique has inherently low penetration depth which gives problems with fundamentally inhomogeneous samples. Also the crystal plates which have to be suitable for use with infra-red spectroscopy are expensive and thus have to be cleaned after each measurement for re-use.
Another known technique is described in U.S. Pat. No. 5,453,252. This makes use of a mesh which defines a plurality of pores. The sample to be analysed is held in the pores by surface tension effects. The technique relies on analysis in transmission and does not produce a film having a reliably uniform thickness.